Measure Customer Experience – But Don’t Over-Engineer


Share on LinkedIn

You have determined for your SME business that you want to improve your customers’ experience.But you do not know how to measure it. And you want tangible results fast. And you want to contain the risk.  After all there is nothing more risky than a big venture that promises results only in the far away future…

You have heard about this Internet of Things thing, beacons, customer journeys, sentiment analysis, and predictive, even intent driven analysis. But this all seems a bit big and daunting.

You ask yourself: How to go ahead? And how do I measure success?

Many companies, especially bigger ones, already have a voice of the customer (VoC) program, which is a good start. Not to be gotten wrong, there are all sorts of challenges with VoC programs. Getting insufficient replies, data that doesn’t really help, data that covers only one single channel, etc. etc.

But then a VoC program is a very good start, if kept simple, and also consistently available across channels. It again is about thinking big while acting small. Key is to

  • Ask the questions immediately in context with the activity you refer to, e.g. on exiting the store or the web site
  • Keep the number of questions for your customers very low and to make answering them extremely easy.
  • Correlate the customer replies with information that you get from your employees.
  • Lastly: Act on the results. Nothing is worse than asking customers and employees for feedback and then doing nothing with it.

Keeping the numbers very low while retaining the ability to get meaningful data is a stretch. There should not be more than 4, in a way that they can be displayed on a mobile phone, ideally without any scrolling. One can do with two questions to get actionable data, but there are two more that should add enough insight to be seriously considered when measuring your customer experience.

And it means simple click or drag answers, e.g. on a sliding scale, or just yes/no.

In a store scenario deliver the questions right after the transaction at the PoS, or on exiting the store. Tie it in to your loyalty/membership program and offer your customers an incentive. This combination offers the possibility of increasing the number of replies, which will improve the accuracy of the measurement. Of course it also increases the importance that you should give your program. It also ties the results to a location. Platforms like Epikonic offer this possibility. You can and should do the same on your e-commerce platform to augment the site log data, and in the up-and-coming chat applications that offer rich communication with brands and that will play more of a role in the future.

However, the results that you get by asking the customers only show one side of the coin. It is equally important to ask your employees.

Regularly. Frequently.

Employees are providing the other side of the measure. If customers are not happy then this is often due to process issues with the (possibly correspondingly disgruntled) personnel that may spend more time than necessary with activities that do not help the sale. Ask the employees that are working on the processes that touch the part of the business that you look at. They are the people on the ground who are experiencing the day-to-day life and they regularly have more insight than HQ personnel. Employee’s answers to two simple questions will help analysing this.

By now you are probably wondering: Which questions, and how to measure?So, here are my recommendations:

Customer questions

  1. Ho would you rate today’s experience? Put this on a scale from 0 – 10. This gives sufficient detail and nicely maps into the common 5 star model. It is also the same scale that the Net Promoter Score uses.
  2. How likely are you to recommend us? Yes, NPS. It is a criticised model but still holds good value. And an experience rating does not make an active supporter.
  3. Did you achieve your goal? This is a Yes/No question and needs to be asked in combination with
  4. What was your goal? Make it a multiple choice question. Good reply candidates are

– Purchase

– Product comparison/information

– Browsing

– Reclamation/Complaint

 Employee questions

  1. What are your biggest pain points?
  2. What change would help you most to improve your ability to help customers?

Put enough emphasis on the first one. It might be the easier one to achieve and also yield considerable results.

Getting this infrastructure in place will bring you onto a good road, especially when combined with already existing data on purchases, web site click throughs, abandoned carts, reactions to campaigns, and more. However, it is not all. From this stable base you can go on to include the sentiments coming in from Facebook, Twitter, Instagram, Pinterest, TripAdvisor, your branded community, blogs, and what not, plus any additional touch point that is there. These will increasingly have the sensors that supply you with a constant stream of pre- and post purchase data.

Including the totally connected individual that we know from Minority Report. But that is not so desirable …

Again: Do not over-engineer the beginnings. Think Big, Act Small, and keep the end in mind, and measure continuously.


  1. Though I heartily disagree that using NPS in any form or fashion will offer much guidance on granular, emotional elements of the customer experience (and you can check my array of CustomerThink blogs on this subject), your questions of employees also miss a fundamental learning and insight opportunity. This is a form of “mirroring” research, where employees are asked how they think customers will rate the experience, and why. There are often profound, and revealing, differences between what experience value employees think they are delivering and what customers actually think and feel..

  2. Thanks for your thoughts, Michael. I agree that none of the questions that I suggest gives much of detail. I also agree that the opinions of the employees are influenced by their own personal context. That is why I would not ask them how they think the customers rate them but what their pain points are (doing their job) and what they think could be done to help them making their job easier. Of course an unstated assumption is that there is a working ‘performance management’ system that supports a customer centric culture. If a business doesn’t want this then improving the customer experience will be a difficult task at best.

    I consider these questionnaires a start, a vehicle to get rolling, with the emphasis on starting simple. Especially small and medium sized retailers need to begin with something simple, for operational reasons as well as for budgetary ones, as they usually operate on slim margins.

    Using this start it is possible to identify how much of a problem exists, and roughly where. If correlating the often existing process data, opinions of the employee, and customer answers we stand a chance of finding and implementing possible, simple, solutions, quick wins.

    Once the discipline of measuring the customer experience is established the process can and should be gradually and iteratively refined, with the ultimate goal of developing the ability of measuring every touch point.

    Be sure that I will read more of your posts. Thanks a lot again!

  3. The challenge with just asking employees about their pain points and what would make their job easier is that, without understanding the context relative to customer value delivery, this information is often too abstract and incomplete. Applying only these employee perceptions might be a weak, off-target application of resources. Mirroring, such as I’ve stated, takes a company much closer to matching up customer and employee points of view, and successfully identifying significant value gaps, Here is a post on the subject from a few years ago:

  4. Have read that post and mainly agree that more data is better. Mirroring is frequently used, also in other areas. The problem with this type of full blown questionnaire is that it is hard to get customers (I am mainly coming from B2C businesses) answering it – and then frequently. Another one is that there is a good risk to find a ‘constant’ gap in self assessment and customer assessment. As said – my approach is just a start to get things rolling. It needs to start simple, else it will with a high likelihood just flounder.

    I do not think that we have a fundamental disagreement.


Please enter your comment!
Please enter your name here